DWO & Tool tip control accuracy

[So not a Guru]

We had a crash on one of our Haas UMC750 table-table 5x machines that required a stream of Haas techs to get working, to what they are saying is correct. I'm less impressed by the accuracy that our simple 5x test program is getting,

The tech is telling us that dynamic work offsets and tool tip control, on 5x work, will never hold tolerances of closer than +/-  0.002" to 0.003".

Is this true? I can see that some machines might be less accurate than others, but this guy is saying it is true no matter what machine is being used.

 

[Guest]

Slight swerve....

No reliable verification, right?

Betting that $$$ would have already paid for itself ....

 

now back to your regularly scheduled HAAS bashing

[Colin Gilchrist]

41 minutes ago, So not a Guru said:

We had a crash on one of our Haas UMC750 table-table 5x machines that required a stream of Haas techs to get working, to what they are saying is correct. I'm less impressed by the accuracy that our simple 5x test program is getting,

The tech is telling us that dynamic work offsets and tool tip control, on 5x work, will never hold tolerances of closer than +/-  0.002" to 0.003".

Is this true? I can see that some machines might be less accurate than others, but this guy is saying it is true no matter what machine is being used.

 

"True no matter what machine" - Absolutely not. There are plenty of Japanese, Swiss, and German machine tool builders that would make a liar out of him.

"True for a Haas?" - Absolutely...

[crazy^millman]

55 minutes ago, So not a Guru said:

We had a crash on one of our Haas UMC750 table-table 5x machines that required a stream of Haas techs to get working, to what they are saying is correct. I'm less impressed by the accuracy that our simple 5x test program is getting,

The tech is telling us that dynamic work offsets and tool tip control, on 5x work, will never hold tolerances of closer than +/-  0.002" to 0.003".

Is this true? I can see that some machines might be less accurate than others, but this guy is saying it is true no matter what machine is being used.

 

Ignorant statement if I every heard one if my entire life. I have seen holding .0001" on a project day in and day out. Yes it was a Matsurra, but I have seen that on Hitachi Seiki, Yasda, Mitsui Seiki and others. We are hitting .0005" on some OKK machines right now without even trying. We could hold .001 on some old SNK machine 15 years ago. You get what you pay for.

[crazy^millman]

Here is where I would look 1st for a Table Table Vertical.

Tied for 1st

Tied for 1st

Still Tied for 1st

solid choice also

I have been around and programmed for every one of these machines and they all will put a HAAS to shame and will hold the tolerances with no issue.

[Watcher]

Now go to the Haas tech and ask him how many machine builders he already worked for. 

Maybe Haas is the only kool-aid he ever had in his entire machine technician life. 

Roeders, Hermle and all brands Ron mentioned are all high accuracy 5 axis machines, all with acceleration rates far higher than a Haas with positional accuracy at least 10x higher as well. 

The Haas technician was unfortunate in his statement. 

[?Mark]

We can adjust parameters on our matsuura 5axis  machines to within .0005 in ALL 5axis fairly easily with a small test program. I bet we could get it tighter than this if we needed to.

 

Not surprised by the haas performance though. There is a GOOD reason they're so much less expensive.

[civiceg]

https://diy.haascnc.com/reference-docs/mrzp-and-dwo#gsc.tab=0

Haas' stance.

You can achieve much better results on a UMC by running Renishaw's Axiset to more accurately find the center point of rotation. I run a 5 sided test program and achieve +/- .0007 on all wall thicknesses without breaking a sweat, using DWO. Machine warm up is also key. 

[So not a Guru]

Thanks guys. I was pretty sure he was mistaken, but I most definitely don't know it all.

To be fair, the guy fixed a machine that 3 other techs couldn't get fixed. And he was, with the one noted exception, very helpful.

[Colin Gilchrist]

And to be fair, the guy just may not know what he is doing very well. I can tell you for DWO on a Haas machine, if you first move your geometry to "close" to where it is on the table, you should be able to get within +-.001 fairly easily. If there are features where "DWO" just won't get you close enough, then use a separate Work Offset for that particular rotation, and then use the Part Probe to get within +-.0003 all day long.

You can make a great part, that is in tolerance with a Haas machine. I've seen people hold bores to within +-.0001 on diameter, and .002 True Position for feature location, but they used the right process for the machine they have. (Adjustable Boring Head, and Probing.) You have to be willing to suit the manufacturing process with the machine you are using. Compare that to a properly calibrated Hermle machine, where you are Interpolating Bores to +-.0001 tolerance, and .0002 True Position, and it starts to open your eyes and mind up a little bit.

I highly recommend that anyone interested in achieving high precision tolerances invest in these two books:

http://mooretool.com/publications.html

Foundations of Mechanical Accuracy is the more important of the two in my opinion. You can learn how much skill and effort it takes to design not just a machine, but the process to build machines, that can achieve tolerances measured in the millionths of an inch. And they were doing this in the 1970's.

[So not a Guru]

5 minutes ago, civicegg said:

You can achieve much better results on a UMC by running Renishaw's Axiset

I'm unfamiliar with this. We have renishaw probes & regularly run Haas' MZRP program that uses a calibrated tooling ball. I'll look into the Axiset process.

Thanks

[civiceg]

Feel free to shoot me any questions you might have. 

[Leon82]

4 hours ago, C^Millman said:

Here is where I would look 1st for a Table Table Vertical.

Tied for 1st

Tied for 1st

Still Tied for 1st

solid choice also

I have been around and programmed for every one of these machines and they all will put a HAAS to shame and will hold the tolerances with no issue.

We have 4 and just got a 330 too.

I ran a part which was 2 pieces. Each was done in a separate 520 and when we bolted them up the true position of the top bore on the top piece, on an 11 degree angle was under .001 and the profile of the cbore face was .0008 on the worst one.

We were worried we would need to cut it assembled. The g43 clearance would have been a half inch. But it came right in cutting each part individually

 

[Leon82]

51 minutes ago, Colin Gilchrist said:

And to be fair, the guy just may not know what he is doing very well. I can tell you for DWO on a Haas machine, if you first move your geometry to "close" to where it is on the table, you should be able to get within +-.001 fairly easily. If there are features where "DWO" just won't get you close enough, then use a separate Work Offset for that particular rotation, and then use the Part Probe to get within +-.0003 all day long.

You can make a great part, that is in tolerance with a Haas machine. I've seen people hold bores to within +-.0001 on diameter, and .002 True Position for feature location, but they used the right process for the machine they have. (Adjustable Boring Head, and Probing.) You have to be willing to suit the manufacturing process with the machine you are using. Compare that to a properly calibrated Hermle machine, where you are Interpolating Bores to +-.0001 tolerance, and .0002 True Position, and it starts to open your eyes and mind up a little bit.

I highly recommend that anyone interested in achieving high precision tolerances invest in these two books:

http://mooretool.com/publications.html

Foundations of Mechanical Accuracy is the more important of the two in my opinion. You can learn how much skill and effort it takes to design not just a machine, but the process to build machines, that can achieve tolerances measured in the millionths of an inch. And they were doing this in the 1970's.

We have a bunch of those books near our Moore's. Its crazy how they could do that precise manually

[Guest]

6 hours ago, So not a Guru said:

The tech is telling us that dynamic work offsets and tool tip control, on 5x work, will never hold tolerances of closer than +/-  0.002" to 0.003"....

That number (.002"-.003") is probably reasonably representative of what the Haas can routinely do in 5-Axis. Using their version of Axi-Set can help that quite a bit, BUT, it is still no way shape or form anywhere near what a Micron, Hermele, GF, Matsuura, Makino, Yasda, Mitsui Seiki, Okuma, and a few others that escape me at the moment are capable of doing.

The tech should probably educate himself a little more about the capabilities of the various High-End MTB's offerings. They don't all share Haas' mechanical AND software limitations. 

 

[Guest]

5 hours ago, Colin Gilchrist said:

"True no matter what machine" - Absolutely not. There are plenty of Japanese, Swiss, and German machine tool builders that would make a liar out of him.

"True for a Haas?" - Absolutely...

Yep.

[crazy^millman]

2 hours ago, Colin Gilchrist said:

I highly recommend that anyone interested in achieving high precision tolerances invest in these two books:

http://mooretool.com/publications.html

Foundations of Mechanical Accuracy is the more important of the two in my opinion. You can learn how much skill and effort it takes to design not just a machine, but the process to build machines, that can achieve tolerances measured in the millionths of an inch. And they were doing this in the 1970's.

i had the shop foreman of Moore move to Jacksonville about 20 something years ago. He worked for me and he was a sharp guy. He knew things about those tight tolerance parts that just blew my mind.

[Greg_J]

What do you guys think of this machine?

https://www.mazakcanada.com/machines/variaxis-i-600/

 

[Zoffen]

I programmed,setup, and ran a i-700 with a palletec system for 3 years and overall i would say its a good machine. It was the first 5 axis i've used so i don't have anything to compare it to. Just watch out if you get the 120tool option as if your tool runs less than 30 seconds you will be waiting on the toolchanger. We ran lang makro grip vices and they worked out great. The machine will move a little as it heatsup/ cools down on some parts i'd have to run a warm-up right before running the part to get consistent parts right off the bat. This was on parts we were trying to hold .001ish true position on features on a 2nd op in a non climate controlled shop. With probing we were able to do this. This cell ran mostly aluminum and the i-700 cut it like butta! The only complaint i had with the machine was mostly due to mazak service. When they installed the machine they had the "Mazacheck"(same as Renishaw Axiset) feature and tried to get it working but no luck. Only about 2 years later were they able to come in and get it working. This was after our maintenance guy had many calls to service and got nowhere and eventually gave up. I had to call the HEAD OF MAZAK SERVICE and politely tell him it was B*LLSH*T they sold us an option that was never fully installed and working even after 2 years!!!! As we were an all mazak shop this was very disappointing to us. The next month they were there with techs from japan to get it working. Funny how if you call the right people stuff happens quick! Overall it was a good machine and made money.

The palletec on the otherhand was a different issue. Software bug causing the system to mismanage pallets!!!! WTF!!!! Good thing the operator was paying attention. This happened a few times and had to deal with when the operator would put the palletchanger in the standby position. I found the software that ran the palletec limited compared to what i feel it should be able to do when compared to other systems. Once working though it did make parts day in and day out.

 

If you any specific questions feel free to ask or PM me.

 

SORRY FOR THE THREAD HIJACK!

[Zoffen]

I have some walkaround and other videos of the i-700 on my youtube if you want to see more:

 

[Greg_J]

Thanks for the reply Zoffen!

I used the Palletec system a few years bad with a Integrex i630 and it worked well, the interface was pretty simple but we weren't doing production at the time.

We don't do any true 5 axis work here but would greatly decrease production time with the "done in one" mindset. I was considering the machine for use with a tombstone and vises set on them to do multiple parts and as many setups in one setup. See following video.

 

[Zoffen]

We mainly were setup for single vise per pallet for most of our work. Its a jobshop so we could never really justify the investment in tombstones and the extra programming time. Typical part runs were 20-50ish parts.

Typical part:

I didn't do much full 5 for any production parts but did make a few personal parts that were full 5. The machine ran pretty good for full 5 as far as i was concerned. Its not a hermle or roeders or makino, its a mazak but i was stoked at the time. Funny how jaded we can get!

For fun:

 

We did make a tombstone specifically for work that never materialized. Fortunately we could use it for other jobs but at 14" wide it was really too large for the machine for general work. A 10" would have been perfect but we never got one before i left. There's a tombstone video on my youtube of that 14" tombstone. When you stack a vise on the side of the tombstone and then throw in some 20xd drills you will run out of Z real quickly

One thing i would invest in if i got that machine would be one of those coolant window spinner things. With 1000psi coolant and 18k you can't see anything sometimes. We did vericut every program but it would have been nice for the operator to see what going on.

 

[Guess_who]

16 hours ago, So not a Guru said:

I'm unfamiliar with this. We have renishaw probes & regularly run Haas' MZRP program that uses a calibrated tooling ball. I'll look into the Axiset process.

Thanks

We just recently purchased Axiset for our Makino Horizontals and our D500's. I highly recommend it. Before this, we had created our own program to set center. I don't think Axiset is more accurate than ours, but it definitely has more options and whistle and bells. All for a reasonable price. 

[gocodeinmotion]

On ‎8‎/‎10‎/‎2017 at 11:24 AM, So not a Guru said:

We had a crash on one of our Haas UMC750 table-table 5x machines that required a stream of Haas techs to get working, to what they are saying is correct. I'm less impressed by the accuracy that our simple 5x test program is getting,

The tech is telling us that dynamic work offsets and tool tip control, on 5x work, will never hold tolerances of closer than +/-  0.002" to 0.003".

Is this true? I can see that some machines might be less accurate than others, but this guy is saying it is true no matter what machine is being used.

 

I haven't read thru this whole thread but there is a way of dialing in your UMC to be more accurate. I've been thru the process as I am an owner too that was plagued with the in accuracy. What control do you nave NGC or legacy?

[gocodeinmotion]

On ‎8‎/‎10‎/‎2017 at 4:14 PM, So not a Guru said:

I'm unfamiliar with this. We have renishaw probes & regularly run Haas' MZRP program that uses a calibrated tooling ball. I'll look into the Axiset process.

Thanks